Proteasome dysfunction triggers activation of SKN-1A/Nrf1 by the aspartic protease DDI-1

Abstract
eLife digest
Introduction
Results
Discussion
Materials and methods
References
Article and author information
Metrics

Abstract

Proteasomes are essential for protein homeostasis in eukaryotes. To preserve cellular function, transcription of proteasome subunit genes is induced in response to proteasome dysfunction caused by pathogen attacks or proteasome inhibitor drugs. In Caenorhabditis elegans, this response requires SKN-1, a transcription factor related to mammalian Nrf1/2. Here, we use comprehensive genetic analyses to identify the pathway required for C. elegans to detect proteasome dysfunction and activate SKN-1. Genes required for SKN-1 activation encode regulators of ER traffic, a peptide N-glycanase, and DDI-1, a conserved aspartic protease. DDI-1 expression is induced by proteasome dysfunction, and we show that DDI-1 is required to cleave and activate an ER-associated isoform of SKN-1. Mammalian Nrf1 is also ER-associated and subject to proteolytic cleavage, suggesting a conserved mechanism of proteasome surveillance. Targeting mammalian DDI1 protease could mitigate effects of proteasome dysfunction in aging and protein aggregation disorders, or increase effectiveness of proteasome inhibitor cancer chemotherapies.

https://doi.org/10.7554/eLife.17721.001

eLife digest

Proteins perform many important roles in cells, but these molecules can become toxic if they are damaged or are no longer needed. A molecular machine called the proteasome destroys ‘unwanted’ proteins in animal and other eukaryotic cells. If the proteasome stops working properly, unwanted proteins start to accumulate and cells respond by increasing the activity of genes that make proteasomes. A protein called SKN-1 is involved in this response and activates the genes that encode proteasome proteins, but it is not understood how SKN-1 “senses” that proteasomes are not working properly.

Here, Lehrbach and Ruvkun used a roundworm called Caenorhabditis elegans to search for new genes that activate SKN-1 when the proteasome’s activity is impaired. The roundworms were genetically engineered to produce a fluorescent protein that indicates when a particular gene needed to make proteasomes is active. Lehrbach and Ruvkun identified some roundworms with mutations that cause the levels of fluorescence to be lower, indicating that SKN-1 was less active in these animals. Further experiments showed that some of these mutations are in genes that encode enzymes called DDI-1 and PNG-1. DDI-1 is able to cut certain proteins, while PNG-1 can remove sugars that are attached to proteins. Therefore, it is likely that these enzymes directly interact with SKN-1 and alter it to activate the genes that produce the proteasome.

More work is now needed to understand the details of how modifying SKN-1 changes its activity in cells. In the future, drugs that target DDI-1 or PNG-1 might be used to treat diseases in which proteasome activity is too high or low, including certain cancers and neurodegenerative diseases.

https://doi.org/10.7554/eLife.17721.002

Introduction

The proteasome is a multi-protein complex responsible for the majority of protein degradation in eukaryotic cells (Tomko and Hochstrasser, 2013). The essential function of the proteasome, and its highly conserved structure and mechanism of proteolysis renders it an attractive target for bacteria and other competitors. Production of small molecule inhibitors and protein virulence factors that target the proteasome by some bacteria and fungi exploits this vulnerability to gain a growth advantage (Fenteany et al., 1995; Groll et al., 2008; Meng et al., 1999). In addition, environmental stresses antagonize the proteasome by causing accumulation of unfolded and aggregated proteins that can form a non-productive inhibitory interaction with proteasomes (Ayyadevara et al., 2015; Deriziotis et al., 2011; Kristiansen et al., 2007; Snyder et al., 2003). Human diseases in which proteasome dysfunction is implicated highlight the importance of maintaining proteasome function in the face of these challenges (Ciechanover and Kwon, 2015; Paul, 2008; Tomko and Hochstrasser, 2013), and it follows that animal cells possess mechanisms to monitor and defend proteasome function.

A conserved response to proteasome disruption is the transcriptional up-regulation of proteasome subunit genes (Fleming, 2002; Meiners et al., 2003; Wójcik and DeMartino, 2002). In mammalian cells members of the Cap’ n’ Collar basic leucine zipper (CnC-bZip) family of stress responsive transcription factors mediate this transcriptional response. Two CnC-bZip franscription factors, Nrf1/NFE2L1 and Nrf2, have similar DNA-binding domains and may regulate an overlapping set of downstream targets. However, only Nrf1 is required for upregulation of proteasome subunits following proteasome disruption, whereas Nrf2 may activate proteasome expression under other circumstances (Arlt et al., 2009; Radhakrishnan et al., 2010; Steffen et al., 2010). The events leading to Nrf1 activation in response to proteasome disruption are complex. In vitro analyses in human and mouse cells indicate that Nrf1 is an endoplasmic reticulum (ER) membrane associated glycoprotein that is constitutively targeted for proteasomal degradation by the ER-associated degradation (ERAD) pathway. Upon proteasome inhibition Nrf1 is stabilized, undergoes deglycosylation and proteolytic cleavage, and localizes to the nucleus (Radhakrishnan et al., 2014; Sha and Goldberg, 2014; Wang, 2006; Zhang and Hayes, 2013; Zhang et al., 2015, 2007, 2014). How processing of Nrf1 is orchestrated, and its significance in responses to proteasome disruption in vivo are not understood.

Upon proteasome disruption, C. elegans induces transcription of proteasome subunit, detoxification, and immune response genes, and animals alter their behavior to avoid their bacterial food source (Li et al., 2011; Melo and Ruvkun, 2012). The transcriptional response to proteasome disruption involves skn-1, which encodes multiple isoforms of a transcription factor with similarities to both Nrf1 and Nrf2 (Blackwell et al., 2015; Li et al., 2011). skn-1 was originally identified for its essential role in embryonic development (Bowerman et al., 1992), but is also required after these early stages for stress responses in a manner analogous to mammalian Nrf1/2 (An and Blackwell, 2003; Oliveira et al., 2009; Paek et al., 2012). SKN-1 binds to the promoters of proteasome subunit genes and mediates their upregulation in response to proteasome disruption, and is required for survival of a mutant with attenuated proteasome function (Keith et al., 2016; Li et al., 2011; Niu et al., 2011). The molecular mechanism that links SKN-1 activation to the detection of proteasome dysfunction has not been established.

Here, we use genetic analysis to uncover the mechanism that couples detection of proteasome defects to these transcriptional responses in C. elegans. We find that an ER-associated isoform of SKN-1 (SKN-1A), is essential for this response. Our genetic data show that the ER-association of this transcription factor normally targets it for poteasomal degradation via ERAD, but is also required for its correct post-translational processing and activation during proteasome dysfunction. After ER-trafficking, our data argues that the PNG-1 peptide N-glycanase removes glycosylation modifications that occur in the ER, and then the DDI-1 aspartic protease cleaves SKN-1A. Each of these steps in SKN-1A processing is essential for the normal response to proteasomal dysfunction. This pathway is essential for compensation of proteasome function under conditions that partially disrupt the proteasome; when compensation is disabled, mild inhibition of the proteasome causes lethal arrest of development. Thus we reveal a vital mechanism of proteasome surveillance and homeostasis in animals.

Results

The aspartic protease DDI-1 and ERAD factors are required for transcriptional responses to proteasome disruption

The proteasome subunit gene rpt-3 is upregulated in a skn-1-dependent manner in response to proteasome disruption (Li et al., 2011). We generated a chromosomally integrated transcriptional reporter in which the rpt-3 promoter drives expression of GFP (rpt-3::gfp). This reporter gene is upregulated in response to drugs such as bortezomib or mutations that cause proteasome dysfunction. To identify the genetic pathways that sense proteasome dysfunction and trigger the activation of SKN-1, we took advantage of a regulatory allele affecting the pbs-5 locus. pbs-5 encodes the C. elegans ortholog of the beta 5 subunit of the 20S proteasome. The pbs-5(mg502) mutation causes constitutive skn-1-dependent activation of rpt-3::gfp expression, but does not otherwise alter fertility or viability (Figure 1—figure supplement 1). Following EMS mutagenesis, we isolated a collection of recessive mutations that suppress the activation of rpt-3::gfp caused by pbs-5(mg502), and identified the causative mutations by whole genome sequencing (Table 1). The collection includes multiple alleles of genes encoding factors required for ERAD. In ERAD, misfolded glycoproteins are retrotranslocated from the ER lumen to the cytoplasm, where they are degraded by the proteasome (Smith et al., 2011). We isolated 3 alleles of sel-1, a gene that encodes the C. elegans orthologue of HRD3/SEL1. HRD3/SEL1 localizes to the ER membrane and recognizes ERAD substrates in the ER (Carvalho et al., 2006; Denic et al., 2006; Gauss et al., 2006), and a single allele of sel-9, which encodes the C. elegans orthologue of TMED2/EMP24, which is also ER-localized and implicated in ER quality control (Copic et al., 2009; Wen and Greenwald, 1999). We also found mutations in png-1, which encodes the C. elegans orthologue of PNG1/NGLY1. After ERAD substrates have been retrotranslocated to the cytoplasm, PNG1/NGLY1 removes N-linked glycans to allow their degradation by the proteasome (Kim et al., 2006; Suzuki et al., 2016). Most strikingly, we isolated six alleles of C01G5.6 (hereafter ddi-1), which encodes the C. elegans orthologue of DDI1 (DNA damage inducible 1). DDI-1 is an aspartic protease, highly conserved throughout eukaryotes (Sirkis et al., 2006). DD1’s function is poorly understood, but it has been implicated in regulation of proteasome function and protein secretion (Kaplun et al., 2005; White et al., 2011).

Table 1

EMS-induced mutations that disrupt rpt-3::gfp activation.

https://doi.org/10.7554/eLife.17721.003

Allele	Affected gene	Effect	Homologues
mg563	C01G5.6	L245F	DDI1. Aspartic protease.
mg555	C01G5.6	C277S
mg544	C01G5.6	G293R
mg542	C01G5.6	R350STOP
mg543	C01G5.6	M244I
mg557	C01G5.6	L334F
mg565	sel-1	G594E	HRD3/SEL1. ER membrane protein, required for ERAD substrate recognition.
mg567	sel-1	A522T
mg547	sel-1	splice site
mg550	sel-9	S140F	EMP24/TMED2. ER membrane protein.
mg561	png-1	G498R	PNG1/NJLY1. Peptide N-glycanase. Removes N-linked glycans during ERAD.
mg564	png-1	splice site

We examined activation of rpt-3::gfp in ERAD and ddi-1 mutant animals following disruption of proteasome function by RNAi of the essential proteasome subunit rpt-5. rpt-5(RNAi) caused larval arrest confirming that all genotypes are similarly susceptible to RNAi. While rpt-5(RNAi) causes robust activation of rpt-3::gfp in wild-type animals, mutants lacking ERAD factors or ddi-1 failed to fully activate rpt-3::gfp (Figure 1a). The requirement for sel-11, which encodes an ER-resident ubiquitin ligase required for ERAD (Smith et al., 2011), supports a general requirement for ERAD in activation of rpt-3::gfp expression. These genes are also required for upregulation of rpt-3::gfp following proteasome disruption by bortezomib (data not shown).

Figure 1 with 1 supplement see all

Download asset Open asset

ER-associated degradation factors and the aspartic protease DDI-1 are required for responses to proteasome disruption.

(a) *rpt-3::gfp* expression following disruption of proteasome function by *rpt-5(RNAi*) in various mutant backgrounds. Scale bars 100 µm. (b) Table showing growth vs. arrest phenotypes of various mutants in the presence of bortezomib or upon *rpn-10* RNAi. For RNAi experiments, L1 animals were incubated for 3 days on indicated RNAi plates, and scored for developmental arrest (+; normal development, Lva; larval arrest). For bortezomib experiments, ~15 L1 animals were incubated for 4 days in liquid cultures containing varying concentrations of bortezomib, and scored for developmental progression. The number (n) of replicate bortezomib experiments performed for each genotype is shown on the right. Each colored rectangle is divided into equal parts to show results from each replicate.

https://doi.org/10.7554/eLife.17721.004

Unlike the wild type, mutants defective in ERAD, or lacking DDI-1, arrested or delayed larval development in the presence of low doses of bortezomib (Figure 1b). We analyzed independently derived alleles of png-1 and ddi-1, indicating that hypersensitivity to proteasome inhibition is unlikely to be a consequence of linked background mutations. png-1 animals consistently showed the most severe defect, and were unable to grow in the presence of very low concentrations of bortezomib. Consistent with their drug sensitivity, mild disruption of proteasome function by RNAi-mediated depletion of the non-essential proteasome subunit RPN-10 causes a synthetic larval lethal phenotype in animals mutant for png-1. The bortezomib sensitivity of ddi-1; sel-11 double mutants was not enhanced compared to that of ddi-1 single mutants, suggesting that ddi-1 and ERAD factors act in the same genetic pathway. We conclude that ERAD and DDI-1 are required for transcriptional upregulation of proteasome subunits and survival during proteasome dysfunction. Given the defective activation of rpt-3::gfp, a direct target of SKN-1, it is likely that upon proteasome disruption, these factors are required to activate SKN-1.

SKN-1A, an ER-associated isoform of SKN-1 mediates proteasome homeostasis

The skn-1 gene generates 3 protein isoforms using alternative transcription start sites (SKN-1A, SKN-1B, and SKN-1C; Figure 2a). The isoforms share an identical C-terminal DNA binding domain, but differ in their N-termini. skn-1(RNAi) targets sequences common to all three transcripts. Both skn-1(RNAi), and the skn-1(zu67) mutation cause defective responses to proteasome dysfunction (NL unpublished). skn-1(zu67) is a nonsense allele affecting an exon shared by SKN-1A and SKN-1C, but that does not affect SKN-1B, suggesting SKN-1A and/or SKN-1C are required. SKN-1C encodes a 61 kD protein expressed specifically in the intestine, and SKN-1A encodes a 71 kD protein expressed in most tissues (An and Blackwell, 2003; Bishop and Guarente, 2007; Staab et al., 2014). SKN-1A differs from SKN-1C solely by the presence of 90 additional amino acids at the N-terminus that includes a predicted transmembrane domain (Figure 2b), and SKN-1A has been found to associate with the ER (Glover-Cutter et al., 2013).

Figure 2 with 1 supplement see all

Download asset Open asset

SKN-1A, a transmembrane-domain-containing isoform of SKN-1 mediates transcriptional responses to proteasome disruption.

(a,b) Schematic of the (a) *skn-1* locus and (b) SKN-1A protein. In (a), the CRISPR-induced *skn-1a*-specific mutations are indicated. (c,d) *rpt-3::gfp* induction in wild type and isoform-specific *skn-1* mutants in (c) the *pbs-5(mg502*) mutant, or (d) *rpt-5(RNAi*). Scale bars 100 µm. (e) Developmental arrest of isoform-specific *skn-1* mutants exposed to mild proteasome disruption by *rpn-10(RNAi*) but not on control RNAi. Scale bar 1 mm. (f) Expression and localization of functional SKN-1A::GFP fusion protein after proteasome disruption by *rpt-5(RNAi*). Apparent GFP signal in control treated animals is background auto-fluorescence. Scale bar 10 µm. (g) No developmental arrest of *skn-1* mutants carrying an isoform-specific *skn-1a::gfp* transgene, and exposed to mild proteasome disruption by *rpn-10(RNAi*). Scale bar 1 mm. (h) Table showing growth vs. arrest phenotypes of *skn-1a* mutants in in the presence of bortezomib. All *skn-1a* alleles are identical in their effect on *skn-1a* coding sequence (G2STOP). Experiments performed identically to those shown in Figure 1b, and data for the wild type from Figure 1 are shown for reference.

https://doi.org/10.7554/eLife.17721.006

We used CRISPR/Cas9 to generate an isoform-specific genetic disruption of SKN-1A, by introducing premature stop codons to the skn-1a specific exons of the skn-1 locus (hereafter referred to as skn-1a mutants). Homozygous skn-1a mutant animals are viable, and under standard conditions show a growth rate and fertility indistinguishable from the wild type. However, skn-1a mutant animals fail to activate rpt-3::gfp in the pbs-5(mg502) mutant background, or upon RNAi of essential proteasome subunit genes, or exposure to bortezomib (Figure 2c,d, data not shown). We note that in these experiments skn-1a mutants failed to activate rpt-3::gfp in all tissues, including the intestine, where SKN-1C is expressed. Consistent with the failure to upregulate rpt-3::gfp, skn-1a mutants show larval lethality when proteasome dysfunction is induced by rpn-10(RNAi) or treatment with a low dose of bortezomib (Figure 2e,h). These skn-1a mutations specifically affect SKN-1A, but leave SKN-1B and SKN-1C unaltered, indicating that SKN-1A is essential for normal responses to proteasome disruption and in the absence of SKN-1A, the other isoforms are not sufficient. A number of stimuli that trigger stabilization and nuclear accumulation of a transgenic SKN-1C::GFP fusion protein are known, but relatively little is known about whether these stimuli also affect SKN-1A (Blackwell et al., 2015). We used miniMos transgenesis (Frøkjær-Jensen et al., 2014) to generate genomically integrated single-copy transgenes that expresses C-terminally GFP-tagged full length SKN-1A (SKN-1A::GFP), and a second C-terminally GFP tagged truncated SKN-1A that lacks the DNA binding domain (SKN-1A[∆DBD]::GFP). When driven by the ubiquitously active rpl-28 promoter, we did not observe accumulation of SKN-1A::GFP or SKN-1A[∆DBD]::GFP, consistent with constitutive degradation of these fusion proteins. Upon disruption of proteasome function, we observed stabilization and nuclear localization of SKN-1A::GFP and SKN-1A[∆DBD]::GFP in many tissues (Figure 2f, Figure 2—figure supplement 1a). For unknown reasons, the SKN-1A[∆DBD]::GFP transgene accumulated to higher levels than the full length transgene (Figure 2—figure supplement 1b). We generated similar transgenes to express tagged full length and truncated SKN-1C, but did not observe any effect of proteasome disruption (data not shown). These data suggest proteasome dysfunction triggers activation of SKN-1A, but not SKN-1C.

We introduced the SKN-1A::GFP transgene into the skn-1a(mg570) and skn-1(zu67) mutant backgrounds. SKN-1A::GFP rescued the maternal effect lethal phenotype of skn-1(zu67). SKN-1A::GFP also restored wild-type resistance to proteasome disruption, as assayed by growth on rpn-10(RNAi) (Figure 3), or growth in the presence of low concentrations of bortezomib (data not shown). This indicates that the SKN-1A::GFP fusion protein is functional, and that SKN-1A::GFP is sufficient for normal responses to proteasome dysfunction even in the absence of SKN-1C (which is disrupted by the zu67 allele). As such, the transmembrane-domain-bearing SKN-1A isoform is necessary and sufficient for responses to proteasome dysfunction.

Figure 3

Download asset Open asset

ERAD is required for constitutive SKN-1A degradation, and for activation of SKN-1A upon proteasome disruption.

(a) Western blot showing expression and post-translational processing of SKN-1A[∆DBD]::GFP in ERAD mutant animals, treated with either solvent control (DMSO) or 5 ug/ml bortezomib. SKN-1A[∆DBD]::GFP is only detected in wild-type animals upon bortezomib exposure; a major band at ~70 kD and a minor band at ~90 kD are detected. In ERAD defective mutants, the ~90 kD band is strongly detected under all conditions, and the ~70 kD band appears only following bortezomib treatment. Actin is used as a loading control. (b) Expression and localization of SKN-1A::GFP in wild type and *sel-1* and *sel-11* ERAD defective mutants after proteasome disruption by *rpt-5(RNAi*). In ERAD defective mutants, SKN-1A::GFP fails to localize to the nucleus. Scale bar 10 µm. (c) Expression and localization of SKN-1A::GFP in wild type and *png-1* mutants after proteasome disruption by *rpt-5(RNAi*). In *png-1* mutants, SKN-1A::GFP is able to localize to the nucleus, although at reduced levels compared to the wild-type. Scale bar 10 µm.

https://doi.org/10.7554/eLife.17721.008

Mutation of ddi-1 does not enhance the sensitivity of skn-1a mutants to bortezomib, suggesting that DDI-1 acts through SKN-1A to promote resistance to proteasome inhibitors (Figure 2h). Additionally removing SEL-11 weakly enhanced the bortezomib sensitivity of ddi-1 skn-1a double mutants, and also caused occasional growth defects even in the absence of proteasome disruption, suggesting that ERAD promotes resistance to proteasome inhibitors largely, but not solely, through regulation of SKN-1A. We examined how ERAD factors regulate SKN-1A using the SKN-1A::GFP transgenes. sel-1 and sel-11 mutants accumulate high levels of SKN-1A[∆DBD]::GFP even in the absence of proteasome inhibitors, showing that SKN-1A is constitutively targeted for proteasomal degradation via ERAD (Figure 3a). Upon proteasome disruption, sel-1 and sel-11 mutants show defects in SKN-1A::GFP nuclear localization consistent with defective release from the ER (Figure 3b). Following proteasome disruption in png-1 mutants SKN-1A::GFP localizes to the nucleus, indicating PNG-1 acts downstream of release from the ER (Figure 3b). Lower levels of SKN-1A::GFP accumulate in the nuclei of png-1 mutants than in the wild type, but this mild effect is unlikely to fully account for the severely defective responses to proteasome inhibition in png-1 mutant animals, suggesting retention of glycosylation modifications normally removed by PNG-1 likely disrupts SKN-1A’s nuclear function. These data suggest that activation of ER-associated and N-glycosylated SKN-1A is required for responses to proteasome dysfunction.

DDI-1 aspartic protease localizes to both nucleus and cytoplasm, and is upregulated upon proteasome disruption

To examine the expression and subcellular localization of the DDI-1 protease, we used miniMos to generate a single copy integrated transgene expressing full length DDI-1 fused to GFP at the N-terminus, under the control of the ddi-1 promoter. The GFP::DDI-1 fusion protein is expressed in most tissues and shows diffuse cytoplasmic and nuclear localization under control conditions, and can rescue a ddi-1 mutant (see below). Following disruption of proteasome function by rpt-5(RNAi), GFP::DDI-1 expression is dramatically induced, and GFP::DDI-1 is enriched in nuclei (Figure 4a). We used CRISPR/Cas9 to modify the ddi-1 locus to incorporate an HA epitope tag near the N-terminus of endogenous DDI-1. Following bortezomib treatment of ddi-1(mg573[HA::ddi-1]) animals, we observed strong upregulation (greater than 10-fold, based on blotting of diluted samples) of the HA-tagged endogenous DDI-1 (Figure 4b). The ddi-1 promoter contains a SKN-1 binding site (Niu et al., 2011). Upregulation of GFP::DDI-1 by rpt-5(RNAi) is greatly reduced in skn-1a(mg570) mutants (Figure 4c), suggesting that DDI-1 upregulation is mostly mediated by SKN-1A. The remaining weaker ddi-1 upregulation in the skn-1a mutant may represent a second skn-1a-independent mechanism that couples DDI-1 levels to proteasome function.

Figure 4

Download asset Open asset

DD1-1 is upregulated upon proteasome disruption.

(a) A functional GFP::DDI-1 fusion protein is strongly induced and localizes to the nucleus upon proteasome disruption by *rpt-5(RNAi*). Scale bar 20 µm. (b) Western blot showing induction of HA-tagged endogenous DDI-1 upon proteasome disruption by bortezomib. Actin is used as a loading control. (c) Induction of GFP::DDI-1 upon proteasome disruption by *rpt-5(RNAi*) is lost in *skn-1a* mutants. Scale bar 20 µm.

https://doi.org/10.7554/eLife.17721.009

DDI-1 is required for proteolytic cleavage of SKN-1A downstream of ER trafficking

The EMS-induced ddi-1 missense alleles that cause failure to activate rpt-3::gfp are clustered within the aspartic protease domain of DDI-1, and affect conserved residues that are thought to form the substrate-binding pocket of the enzyme (Figure 5a,b), suggesting that the protease activity of DDI-1 is required (Sirkis et al., 2006). We used CRISPR/Cas9 mutagenesis to generate a protease dead mutant containing two amino acid substitutions at conserved residues of the catalytic motif, including the aspartic acid residue that forms the active site (D261N, G263A). We additionally isolated a CRISPR-induced deletion that deletes most of the aspartic protease domain and introduces a frameshift, which we presume to be a null allele. Both mutations cause a similar, strong defect in rpt-3::gfp activation by the pbs-5(mg502) mutant, or upon proteasome RNAi, and cause a similar sensitivity to bortezomib (Figure 5c, data not shown).

Figure 5 with 1 supplement see all

Download asset Open asset

The DDI-1 aspartic protease is required for proteolytic activation of SKN-1A.

(a) Schematic of the DDI-1 protein showing residues affected by EMS-induced loss of function alleles. (b) Multiple alignment of the aspartic protease domain of DDI-1. Red text indicates residues affected by mutations that disrupt *rpt-3::gfp* activation, blue text indicates the DS/TGAQ catalytic motif. (c) *rpt-3::gfp* induction in the *pbs-5(mg502)* mutant background (left panel), compared to animals carrying a *ddi-1* deletion (middle panel), or a point mutation affecting the catalytic motif (right panel). Scale bars 100 µm. (d) Localization of SKN-1A::GFP in wild-type and *ddi-1* mutant animals following disruption of proteasome function by *rpt-5(RNAi*). SKN-1A::GFP nuclear localization is intact in the *ddi-1* mutant. In some gut cells of *ddi-1* mutant animals SKN-1A::GFP localizes to puncta that are not detected in the wild type. Scale bars 10 µm. (e) Localization of SKN-1A[∆DBD]::GFP in gut nuclei of wild-type and *ddi-1* mutant animals. Nuclear foci of SKN-1A[∆DBD]::GFP are found in *ddi-1* mutants, but not wild type. (f) Western blot showing expression and processing of SKN-1A[∆DBD]::GFP in *ddi-1* mutant animals, treated with either solvent control (DMSO) or 5 ug/ml bortezomib, and blotted for GFP. In the *ddi-1* mutant animals, the major band detected is ~30 kD larger than in the wild type. (g) Western blot showing expression and processing of HA::SKN-1A:GFP in *ddi-1* mutants animals, treated with either solvent control (DMSO) or 5 ug/ml bortezomib, and blotted for HA. In the wild type, a ~20 kD band is detected in animals exposed to bortezomib. In *ddi-1* mutants this low molecular weight fragment is absent, and a ~110 kD band is detected. In (f) and (g) *ddi-1* mutations were *ddi-1(mg571*)[deletion] or *ddi-1(mg572*)[active site] and actin is used as a loading control.

https://doi.org/10.7554/eLife.17721.010

S. cervisiae Ddi1 contains an N-terminal ubiquitin-like (UBL) domain and a C-terminal ubiquitin-associated (UBA) domain, but these domains are not detected by standard protein sequence comparisons with C. elegans DDI-1. To address the possibility that UBL or UBA domains with highly divergent sequence may be present in DDI-1, we generated N-terminally truncated (∆N), and C-terminally truncated (∆C) gfp::ddi-1 transgenes. We tested their ability to rescue the bortezomib sensitivity phenotype of ddi-1(mg571) alongside wild-type gfp::ddi-1 and an aspartic protease active site (D261N) mutant. The active site mutation abolished rescue by the gfp::ddi-1 transgene, whereas the ∆N and ∆C truncated transgenes restored bortezomib sensitivity to near wild-type levels (Figure 5—figure supplement 1). These data are consistent with the lack of conservation of the UBL and UBA domains of DDI-1, and confirm the essential role of DDI-1 aspartic protease activity.

In animals lacking DDI-1, SKN-1A::GFP localizes at normal levels to the nucleus upon proteasome disruption by rpt-5(RNAi), suggesting that DDI-1 regulates SKN-1A function after nuclear localization of the transcription factor (Figure 5d). We noticed that SKN-1A::GFP occasionally showed abnormal localization within gut nuclei of ddi-1 mutants, accumulating in highly fluorescent puncta. We observed this defect for both SKN-1A::GFP and SKN-1A[∆DBD]::GFP, indicating that the DBD of SKN-1A is not required for this mis-localization (Figure 5e).

As in the wild type, SKN-1A[∆DBD]::GFP does not accumulate in the absence of proteasome disruption in ddi-1 mutants, indicating that the DDI-1 peptidase does not participate in constitutive degradation of SKN-1A by the proteasome (Figure 5f). SKN-1A[∆DBD]::GFP accumulates to similar levels upon proteasome disruption by bortezomib in wild-type and ddi-1 mutants, but in ddi-1 mutants is ~20 kD larger than in the wild type, and approximates the expected size of SKN-1A[∆DBD]::GFP. To test whether these differences reflect DDI-1-dependent proteolytic processing of SKN-1A, we generated a transgene that expresses full length SKN-1A with an N-terminal HA tag and a C-terminal GFP tag (HA::SKN-1A::GFP). The expression, localization and rescue activity of the dually tagged fusion protein is indistinguishable from that of the full length SKN-1A::GFP transgene. In wild-type animals carrying the HA::SKN-1A::GFP transgene, Western blotting for the HA tag reveals a ~20 kD band that accumulates specifically upon proteasome disruption by bortezomib treatment. In ddi-1 deletion or active site mutants, a ~110 kD protein accumulates upon proteasome disruption, equivalent in size to full-length HA::SKN-1::GFP (Figure 5g). As such, SKN-1A is cleaved at a position approximately 20 kD from the N-terminus, and the protease active site of DDI-1 is required for this cleavage.

In sel-1 and sel-11 mutants, unprocessed HA::SKN-1A::GFP is present in both control and bortezomib treated animals, with some processed HA::SKN-1A::GFP appearing upon bortezomib treatment (Figure 6a). In png-1 mutant animals HA::SKN-1A::GFP is incompletely processed following proteasome disruption (Figure 6b), suggesting that DDI-1 dependent cleavage of SKN-1A normally occurs after ER trafficking and deglycosylation. Further, in wild-type animals, an HA::SKN-1A::GFP transgene that lacks the putative transmembrane domain is not targeted for constitutive degradation, and is not subject to detectable proteolytic cleavage upon proteasome disruption (HA::SKN-1A[∆TM]::GFP; Figure 6c). Although expressed, HA::SKN-1A[∆TM]::GFP is unable to rescue the bortezomib sensitivity of skn-1a mutants (data not shown). These results indicate that ER trafficking is needed to target SKN-1A for later cleavage and activation by ddi-1.

Figure 6 with 1 supplement see all

Download asset Open asset

Proteolytic processing of SKN-1A occurs downstream of ER trafficking.

(a) Western blot showing expression and processing of HA::SKN-1A::GFP upon proteasome disruption by bortezomib in *sel-1* and *sel-11* mutant animals. ERAD-defective mutants accumulate similar levels of uncleaved (~110 kD) HA::SKN-1A::GFP in both absence and presence of bortezomib. A relatively small amount of the ~20 kD cleavage product accumulates only upon exposure to bortezomib. (b) Western blots comparing expression and processing of HA::SKN-1A::GFP upon proteasome disruption by bortezomib in wild-type and *png-1* mutant animals. Both cleaved (~20 kD) and uncleaved (smear ~100-150 kD) HA::SKN-1A::GFP accumulates upon bortezomib treatment in *png-1* mutants. (c) Western blot showing expression and processing of HA::SKN-1A[∆TM]::GFP upon proteasome disruption by bortezomib, in otherwise wild-type animals. Uncleaved (~100 kD) HA::SKN-1A[∆TM]::GFP is detected at similar levels under both conditions suggesting the protein is not subject to proteasomal degradation, and a low molecular weight cleavage product is not detected under either condition. For each experiment, mixed stage cultures were treated with either solvent control (-), or 5 ug/ml bortezomib (+) prior to collection for SDS-PAGE. HA::SKN-1A::GFP/ HA::SKN-1A[∆TM]::GFP is detected by anti-HA antibodies, Actin is used as a loading control.

https://doi.org/10.7554/eLife.17721.012

Discussion

Our genetic screen for mutants that fail to activate SKN-1 and dissection of the isoform-specific role of SKN-1A reveals the molecular details of proteasome surveillance. We show that SKN-1A is an ER associated protein that is normally targeted by the ERAD pathway for proteasomal degradation in the cytoplasm. Mutations affecting ERAD genes sel-1/SEL1/HRD3 and sel-11/HRD1, stabilize SKN-1A, but also disrupt localization of SKN-1A following proteasome disruption, due to a failure to efficiently release SKN-1A from the ER. Our data argues that following release from the ER, SKN-1A must be deglycosylated by PNG-1 and cleaved by DDI-1 to become fully active Figure 6—figure supplement 1.

The bortezomib sensitivity defect of png-1 mutants is similar in strength to skn-1a mutants, and both png-1 and skn-1a mutations are synthetic lethal with rpn-10(RNAi). These similarities suggest that SKN-1A activity may be completely abolished in the absence of PNG-1; likely as a result of failure to deglycosylate SKN-1A following its retrotranslocation from the ER, although we cannot rule out the possibility that deglycosylation of other proteins contributes indirectly to SKN-1A function. Surprisingly, we were unable to generate png-1; skn-1a double mutants, apparently due to lethality of double mutant embryos (NL unpublished). This indicates that in png-1 mutant animals, SKN-1A (likely in a glycosylated state) retains a function that is essential for development. This suggests that SKN-1A is not only important for proteasome homeostasis, but is also important for cellular homeostasis upon disruption of glycoprotein metabolism. Consistently, skn-1 mutants are hypersensitive to tunicamycin, an inhibitor of protein glycosylation (Glover-Cutter et al., 2013).

NGLY1, the human PNG-1 orthologue, plays important roles in human development; NGLY1 deficiency, a recently described genetic disorder of protein deglycosylation, is caused by mutations at the NGLY1 locus (Enns et al., 2014). Failure to deglycosylate Nrf1, and consequent defects in proteasome homeostasis, likely contributes to the symptoms associated with NGLY1 deficiency. As such our work identifies a pathway that may be targeted for treatment of NGLY1 deficiency - genetic screens for suppressors of the defective proteasome gene regulation of png-1 mutants could indicate targets for drug development.

The expression of DDI-1 is dramatically responsive to proteasome inhibition, indicating that its synthesis or stability is coupled to surveillance of proteasome dysfunction. ChIP analysis of SKN-1 shows binding at the promoter element of the DDI-1 operon (Niu et al., 2011), suggesting that DDI-1 upregulation may occur via SKN-1 mediated transcriptional regulation, and we observed a defect in the upregulation of GFP::DDI-1 following proteasome disruption in skn-1a mutants. This suggests a positive feedback loop, wherein SKN-1A upregulates DDI-1, and DDI-1 promotes SKN-1A activation. Positive feedback may ensure a timely and robust response to proteasome inhibition.

Mutation of ddi-1 causes defective regulation of rpt-3::gfp and increases sensitivity of C. elegans to proteasome inhibitors. SKN-1A is cleaved at a site ~20 kD from the N-terminus, and ddi-1 is required for this cleavage. The requirement for its catalytic function strongly suggests that DDI-1 is the enzyme directly responsible for cleavage of SKN-1A, although we cannot rule out the possibility that DDI-1 acts upstream of an as-yet unidentified second protease. There are precedents for cascades of proteases, for example caspases in apoptosis, complement cascades in immunology, and thrombin cascades in blood clotting. Our genetic screen for proteasome surveillance defective mutants isolated six independent ddi-1 alleles, but as yet no alleles of any other genes that encode proteases. This argues that either DDI-1 is the only protease in the pathway, or that any other proteases function redundantly or have other essential functions.

Uncleaved SKN-1A localizes to the nucleus in ddi-1 mutants, so cleavage of SKN-1A is not essential for its nuclear localization, and SKN-1A cleavage may occur either in the nucleus or cytoplasm. Given that GFP::DDI-1 is largely (but not exclusively) nuclear under conditions of proteasome disruption, we speculate that DDI-1 cleavage of SKN-1A is nuclear. In either case, unusually for a membrane-associated transcription factor, SKN-1A is released from the ER by a mechanism that does not require proteolytic cleavage. DDI-1-dependent cleavage therefore activates SKN-1A by some other mechanism(s) downstream of ER release. Cleavage of SKN-1A may be required to remove domains in the N-terminus that interfere with its normal function in the nucleus. For example, retention of the hydrophobic transmembrane domain may be disruptive once the protein has been extracted from the ER membrane.

Mutant SKN-1A lacking the transmembrane domain is not subject to proteasomal degradation, and is not cleaved by DDI-1. So, in addition to serving to link SKN-1A levels to proteasome function via ERAD, ER trafficking of SKN-1A is important for subsequent DDI-1-dependent activation. The bortezomib sensitivity of skn-1a mutants (or skn-1a mutants carrying the transmembrane domain-lacking transgene) is more severe than that of ddi-1 and ERAD mutants, so ER-association must also promote SKN-1A activation by additional mechanisms. As well as proteasome disruption, skn-1 is implicated in responses to several endocrine and environmental stimuli (Blackwell et al., 2015). Modifications such as glycosylation that SKN-1A acquires in the ER may tailor its activity to respond to proteasome dysfunction, identifying these modifications and how they are regulated will be of interest.

S. cerevisiae Ddi1 contains an N-terminal UBL domain and a C-terminal UBA domain, This domain architecture is typical of extraproteasomal ubiquitin receptors, which play a role in recruiting ubiquinated proteins to the proteasome (Tomko and Hochstrasser, 2013). S. cerevisiae Ddi1 binds to both ubiquitin and the proteasome, and participates in the degradation of some proteasome substrates (Bertolaet et al., 2001; Gomez et al., 2011; Kaplun et al., 2005; Nowicka et al., 2015), and synthetic genetic interactions with extraproteasomal ubiquitin receptor and proteasome subunit mutants supports a role for Ddi1 in proteasome function (Costanzo et al., 2010; Díaz-Martínez et al., 2006). Although the aspartic protease domain is highly conserved, DDI-1 in C. elegans and related nematodes apparently lacks both UBL and UBA domains, and the UBA domain is not found in mammalian Ddi1 orthologues, so it remains unclear whether Ddi1 orthologues function as extraproteasomal ubiquitin receptors in animals (Nowicka et al., 2015). The effect of ddi-1 on development upon proteasome disruption by bortezomib is entirely dependent on skn-1a, indicating that DDI-1 promotes resistance to proteasome disruption via SKN-1A, rather than a general effect on proteasome function. Regardless, it will be of interest to determine whether DDI-1 binds to proteasomes and/or ubiquitin, and whether this affects its function in SKN-1A activation.

Activation of the mammalian SKN-1 homologue Nrf1 involves both deglycosylation and proteolytic cleavage, but the enzymes responsible are not known (Radhakrishnan et al., 2014; Zhang et al., 2015). However, a large scale screen for gene inactivations that render cells more sensitive to proteasome inhibitors supports a model that human DDI1 protease also processes Nrf1: DDI2 (one of two human orthologues of DDI-1) and Nrf1 were highly ranked hits in this screen that identified hundreds of gene inactivations that increase sensitivity of multiple myeloma cells to proteasome inhibitors (Acosta-Alvear et al., 2015). This suggests that DDI2 is required to cleave and activate Nrf1 in human cells. The site at which Nrf1 is cleaved during proteasome dysfunction has been identified (Radhakrishnan et al., 2014), but the primary sequence of this site is not conserved in SKN-1A. Comparisons of SKN-1A with its nematode orthologues reveals conservation at positions consistent with the ~20 kD cleavage product we have observed (NL unpublished). It is possible that DDI-1 and its substrate(s) have divergently co-evolved in different lineages. Thus, we suggest that DDI-1 and SKN-1A are core components of a conserved mechanism of proteasome surveillance in animals.

Here we have shown that correct post-translational processing of SKN-1A is essential for development if proteasome function is disrupted. Deregulated proteasome function a feature of aging and age-related disease (Saez and Vilchez, 2014; Taylor and Dillin, 2011). skn-1 is a critical genetic regulator of longevity, and controls lifespan in part through regulation of proteasome function (Blackwell et al., 2015; Steinbaugh et al., 2015) As such, the SKN-1A processing pathway described here suggests the mechanism that links SKN-1/Nrf to proteasome function and longevity.

Proteasome inhibitors are important drugs in the treatment of multiple myeloma, but relapse and emergence of drug resistant tumors remains a challenge (Dou and Zonder, 2014). Nrf1 promotes survival of cancerous cells treated with proteasome inhibitors, and activation of this pathway might mediate resistance (Acosta-Alvear et al., 2015; Radhakrishnan et al., 2010; Steffen et al., 2010). Blocking the activation of proteasome subunit gene expression by Nrf1 has been proposed as a potential strategy to improve effectiveness of proteasome inhibitors in cancer treatment. The conserved SKN-1A/Nrf1 processing factors we have identified, particularly DDI-1, are ideal targets for such an approach.

Materials and methods

C. elegans maintenance and genetics

Request a detailed protocol

C. elegans were maintained on standard media at 20°C and fed E. coli OP50. A list of strains used in this study is provided in Supplementary Table 1. Mutagenesis was performed by treatment of L4 animals in 47 mM EMS for 4 hr at 20°C. RNAi was performed as described in Kamath and Ahringer (2003). The mgIs72[rpt-3::gfp] integrated transgene was generated from sEx15003 (Hunt-Newbury et al., 2007), using EMS mutagenesis to induce integration of the extrachromosomal array. Some strains were provided by the CGC, which is funded by NIH Office of Research Infrastructure Programs (P40 OD010440). sel-11(tm1743) was kindly provided by Shohei Mitani. png-1(ok1654) was generated by the C. elegans Gene Knockout Project at the Oklahoma Medical Research Foundation, part of the International C. elegans Gene Knockout Consortium.

Identification of EMS induced mutations by whole genome sequencing

Request a detailed protocol

Genomic DNA was prepared using the Gentra Puregene Tissue kit (Qiagen, #158689) according to the manufacturer’s instructions. Genomic DNA libraries were prepared using the NEBNext genomic DNA library construction kit (New England Biololabs, #E6040), and sequenced on a Illumina Hiseq instrument. Deep sequencing reads were analyzed using Cloudmap (Minevich et al., 2012).

Following deep sequencing analysis, a number of criteria were taken into account to identify the causative alleles, as shown in Supplemental file 1. In many cases, the causative alleles were strongly suggested by the identification of multiple independent alleles for a given gene. Even for those genes only identified by a single allele, the strong functional connection with other independently mutated genes suggests that they are causative (e.g. isolation of multiple alleles of the sel gene class). We also obtained genetic linkage data supporting these assignments for most alleles. For most of the mutants considered, deep sequencing was performed using a DNA from a pool of 20–50 mutant F2s generated by outcrossing the original mutant strain to the parental (non-mutagenised) background, which allowed us to use Cloudmap variant discovery mapping to identify the genetic linkage of the causative allele, or linkage was confirmed by testing linkage in crosses with strains carrying precisely mapped miniMos insertions. For ddi-1 and png-1, we confirmed that disruption by an independent means (with an independently derived allele) has the same effect on rpt-3::gfp expression as the EMS-induced mutation.

Identification of the pbs-5(mg502) mutation

Request a detailed protocol

The mg502 allele was isolated in an EMS mutagenesis screen in which mgIs72[rpt-3::gfp] animals were screened for recessive mutations causing constitutive activation of GFP expression. The mutation was identified as described above. The pbs-5(mg502) lesion is a 122bp deletion in the promoter of CEOP1752. This operon consists of K05C4.2 and pbs-5. Animals carrying this mutation show constitutive activation of rpt-3::gfp, but have normal growth and fertility under control conditions.

Genome modification by CRISPR/Cas9

Request a detailed protocol

Guide RNAs were selected by searching the desired genomic interval for ‘NNNNNNNNNNNNNNNNNNRRNGG’, using Ape DNA sequence editing software (http://biologylabs.utah.edu/jorgensen/wayned/ape/). All guide RNA constructs were generated by Q5 site directed mutagenesis as described (Dickinson et al., 2013). Repair template oligos were designed as described (Paix et al., 2014; Ward, 2015). Injections were performed using the editing of pha-1 (to restore e2123ts) or dpy-10 (to generate cn64 rollers) as phenotypic co-CRISPR markers (Arribere et al., 2014; Ward, 2015). Injection mixes contained 60 ng/ul each of the co-CRISPR and gene of interest targeting Guide RNA/Cas9 construct, and 50 ng/ul each of the co-CRISPR and gene of interest repair oligos. Guide RNA and homologous repair template sequences are listed in Supplemental file 1.

Transgenesis

Request a detailed protocol

Cloning was performed by isothermal/Gibson assembly (Gibson et al., 2009). All plasmids used for transgenesis are listed in Supplemental file 1. All miniMos constructs were assembled in pNL43, a modified version of pCFJ909 containing the pBluescript MCS, and are described in more detail below. MiniMos transgenic animals were isolated as described, using unc-119 rescue to select transformants (Frøkjaer-Jensen et al., 2014).

MiniMos constructs

skn-1 constructs

Request a detailed protocol

To generate constructs that express specific SKN-1 isoforms, we cloned skn-1c and skn-1a ‘minigenes’. The skn-1c minigene was assembled from 3 genomic fragments from the skn-1 locus: (1) skn-1c exon 1, (2) skn-1c exons 2 & 3, (3) skn-1c exons 4, 5 & 6. These three fragments were fused in frame to generate a skn-1c coding sequence lacking the two largest introns that may contain regulatory information. The skn-1a minigene was assembled by fusing two additional fragments, corresponding to the first two exons of skn-1a, to the 5’ end of the skn-1c minigene, eliminating the large intron that contains the presumed skn-1c promoter. These minigenes were then fused in frame with an N-terminal HA tag and/or a C-terminal GFP tag, and inserted into pNL43 with the rpl-28 promoter (605 bp immediately upstream of the rpl-28 start codon) and tbb-2 3’UTR (376 bp immediately downstream of the tbb-2 stop codon). skn-1a[∆DBD]::gfp and skn-1c[∆DBD]::gfp were generated by deleting sequence corresponding to the final 94 amino acids from the each construct (i.e. amino acids 530–623 of SKN-1A and amino acids 440–533 of SKN-1C) by isothermal assembly using appropriate fragments of the respective minigenes. skn-1a[∆TM]::gfp was generated by deleting sequence corresponding to amino acids 39–59 of SKN-1A, by isothermal assembly of appropriate fragments of the skn-1a minigene.

ddi-1 constructs

Request a detailed protocol

The genomic C01G5.6/ddi-1 coding sequence was fused in frame with GFP at the N-terminus. The gfp::ddi-1 fragment was inserted into pNL43 with the ddi-1 promoter (803 bp immediately upstream of the start codon), and the tbb-2 3’UTR. The gfp::ddi-1[D261N] construct was generated by site-directed mutagenesis, and the N- and C-terminal truncation constructs were generated by isothermal assembly using appropriate fragments of the ddi-1 genomic coding sequence.

Microscopy

Request a detailed protocol

Low magnification bright field and GFP fluorescence images (those showing larval growth and rpt-3::gfp expression) were collected using a Zeiss AxioZoom V16, equipped with a Hammamatsu Orca flash 4.0 digital camera camera, and using Axiovision software. High magnification differential interference contrast (DIC) and GFP fluorescence images (those showing SKN-1A::GFP and GFP::DDI-1 expression) were collected using a Zeiss Axio Image Z1 microscope, equipped with a Zeiss AxioCam HRc digital camera, and using Axiovision software. Images were processed using ImageJ software. For all fluorescence images, any images shown within the same figure panel were collected together using the same exposure time and then processed identically in ImageJ.

Bortezomib sensitivity assays

Request a detailed protocol

Bortezomib sensitivity was assessed by the ability of L1 animals to develop in the presence of a range of concentrations of bortezomib (LC Laboratories, #B1408). The assays were carried out in liquid culture in ½ area 96 well plates (Corning, #3882). Each well contained a total volume of 35uL. We mixed ~15 L1 larvae with concentrated E. coli OP50 suspended in S-basal (equivalent to bacteria from ~200 uL of saturated LB culture), supplemented with 50 ug/ml Carbenicillin, and the desired concentration of bortezomib. All treatment conditions contained 0.01% DMSO. The plates were sealed with Breathe-Easy adhesive film (Diversified Biotech, #9123–6100). The liquid cultures were incubated for 4 days at 20°C and then C. elegans growth was manually scored under a dissecting microscope. Growth was scored into three categories: (1) Normal - indistinguishable from wild type grown in DMSO control, most animals reached adulthood; (2) Delayed development - most animals are L3 or L4 larvae; (3) Larval arrest/lethal - all animals are L1 or L2 larvae. For each genotype all conditions were tested in at least 2 replicate experiments.

Western blot following bortezomib treatment

Request a detailed protocol

Drug treatments were performed in liquid culture in 6-well tissue culture plates. In each well we mixed C. elegans suspended in S-basal (~1000–2000 worms collected from a mixed stage culture grown at 20°C on NGM agar plates) and E. coli OP50 in S-Basal (equivalent to E. coli from ~4 mL saturated LB culture), supplemented with 50 ug/ml Carbenicillin, and the desired concentration of bortezomib, and made the mixture up to a final volume of 700 ul. All wells contained 0.01% DMSO. The tissue culture plates were sealed with Breathe-Easy adhesive film and incubated at 20°C for 7–9 hr. After the treatment the animals were collected to 1.5 ml microcentrifuge tubes, washed twice in PBS to remove bacteria and the worm pellet was snap frozen in liquid nitrogen and stored at −80°C. The worm pellet was briefly thawed on ice, mixed with an equal volume of 2x Sample buffer (20% glycerol, 120 mM Tris pH 6.8, 4% SDS, 0.1 mg/ml bromophenol blue, 5% beta mercaptoethanol), heated to 95°C for 10 min, and centrifuged at 16,000g for 10 min to pellet debris. SDS-PAGE and western blotting was performed using NuPAGE apparatus, 4–12% polyacrylamide Bis-Tris pre-cast gels (Invitrogen, #NP0321) and nitrocellulose membranes (Invitrogen, #LC2000) according to the manufacturer’s instructions. The following antibodies were used: mouse anti-GFP (Roche; #11814460001); HRP-conjugated mouse anti-HA (Roche, # 12013819001), mouse anti-Actin (Abcam; #3280).

Multiple alignment of protein sequences

Request a detailed protocol

Multiple alignment was performed using Clustal Omega (www.ebi.ac.uk/tools/clustalo).

References

1. Acosta-Alvear D
2. Cho MY
3. Wild T
4. Buchholz TJ
5. Lerner AG
6. Simakova O
7. Hahn J
8. Korde N
9. Landgren O
10. Maric I
11. Choudhary C
12. Walter P
13. Weissman JS
14. Kampmann M
(2015) Paradoxical resistance of multiple myeloma to proteasome inhibitors by decreased levels of 19S proteasomal subunits
eLife 4:e08153.

https://doi.org/10.7554/eLife.08153
- Google Scholar
1. An JH
2. Blackwell TK
(2003) SKN-1 links C. elegans mesendodermal specification to a conserved oxidative stress response
Genes & Development 17:1882–1893.

https://doi.org/10.1101/gad.1107803
- Google Scholar
1. Arlt A
2. Bauer I
3. Schafmayer C
4. Tepel J
5. Müerköster SS
6. Brosch M
7. Röder C
8. Kalthoff H
9. Hampe J
10. Moyer MP
11. Fölsch UR
12. Schäfer H
(2009) Increased proteasome subunit protein expression and proteasome activity in colon cancer relate to an enhanced activation of nuclear factor E2-related factor 2 (Nrf2)
Oncogene 28:3983–3996.

https://doi.org/10.1038/onc.2009.264
- Google Scholar
1. Arribere JA
2. Bell RT
3. Fu BX
4. Artiles KL
5. Hartman PS
6. Fire AZ
(2014) Efficient marker-free recovery of custom genetic modifications with CRISPR/Cas9 in Caenorhabditis elegans
Genetics 198:837–846.

https://doi.org/10.1534/genetics.114.169730
- Google Scholar
(2015) Proteins in aggregates functionally impact multiple neurodegenerative disease models by forming proteasome-blocking complexes
Aging Cell 14:35–48.

https://doi.org/10.1111/acel.12296
- Google Scholar
1. Bertolaet BL
2. Clarke DJ
3. Wolff M
4. Watson MH
5. Henze M
6. Divita G
7. Reed SI
(2001) UBA domains of DNA damage-inducible proteins interact with ubiquitin
Nature Structural Biology 8:417–422.

https://doi.org/10.1038/87575
- Google Scholar
1. Bishop NA
2. Guarente L
(2007) Two neurons mediate diet-restriction-induced longevity in C. elegans
Nature 447:545–549.

https://doi.org/10.1038/nature05904
- Google Scholar
(2015) SKN-1/Nrf, stress responses, and aging in Caenorhabditis elegans
Free Radical Biology and Medicine 88:290–301.

https://doi.org/10.1016/j.freeradbiomed.2015.06.008
- Google Scholar
(1992) skn-1, a maternally expressed gene required to specify the fate of ventral blastomeres in the early C. elegans embryo
Cell 68:1061–1075.

https://doi.org/10.1016/0092-8674(92)90078-Q
- Google Scholar
(2006) Distinct ubiquitin-ligase complexes define convergent pathways for the degradation of ER proteins
Cell 126:361–373.

https://doi.org/10.1016/j.cell.2006.05.043
- Google Scholar
1. Ciechanover A
2. Kwon YT
(2015) Degradation of misfolded proteins in neurodegenerative diseases: therapeutic targets and strategies
Experimental & Molecular Medicine 47:e147.

https://doi.org/10.1038/emm.2014.117
- Google Scholar
1. Copic A
2. Dorrington M
3. Pagant S
4. Barry J
5. Lee MC
6. Singh I
7. Hartman JL
8. Miller EA
(2009) Genomewide analysis reveals novel pathways affecting endoplasmic reticulum homeostasis, protein modification and quality control
Genetics 182:757–769.

https://doi.org/10.1534/genetics.109.101105
- Google Scholar
1. Costanzo M
2. Baryshnikova A
3. Bellay J
4. Kim Y
5. Spear ED
6. Sevier CS
7. Ding H
8. Koh JL
9. Toufighi K
10. Mostafavi S
11. Prinz J
12. St Onge RP
13. VanderSluis B
14. Makhnevych T
15. Vizeacoumar FJ
16. Alizadeh S
17. Bahr S
18. Brost RL
19. Chen Y
20. Cokol M
21. Deshpande R
22. Li Z
23. Lin ZY
24. Liang W
25. Marback M
26. Paw J
27. San Luis BJ
28. Shuteriqi E
29. Tong AH
30. van Dyk N
31. Wallace IM
32. Whitney JA
33. Weirauch MT
34. Zhong G
35. Zhu H
36. Houry WA
37. Brudno M
38. Ragibizadeh S
39. Papp B
40. Pál C
41. Roth FP
42. Giaever G
43. Nislow C
44. Troyanskaya OG
45. Bussey H
46. Bader GD
47. Gingras AC
48. Morris QD
49. Kim PM
50. Kaiser CA
51. Myers CL
52. Andrews BJ
53. Boone C
(2010) The genetic landscape of a cell
Science 327:425–431.

https://doi.org/10.1126/science.1180823
- Google Scholar
(2006) A luminal surveillance complex that selects misfolded glycoproteins for ER-associated degradation
Cell 126:349–359.

https://doi.org/10.1016/j.cell.2006.05.045
- Google Scholar
1. Deriziotis P
2. André R
3. Smith DM
4. Goold R
5. Kinghorn KJ
6. Kristiansen M
7. Nathan JA
8. Rosenzweig R
9. Krutauz D
10. Glickman MH
11. Collinge J
12. Goldberg AL
13. Tabrizi SJ
(2011) Misfolded PrP impairs the UPS by interaction with the 20S proteasome and inhibition of substrate entry
EMBO Journal 30:3065–3077.

https://doi.org/10.1038/emboj.2011.224
- Google Scholar
(2013) Engineering the Caenorhabditis elegans genome using Cas9-triggered homologous recombination
Nature Methods 10:1028–1034.

https://doi.org/10.1038/nmeth.2641
- Google Scholar
1. Dou QP
2. Zonder JA
(2014) Overview of proteasome inhibitor-based anti-cancer therapies: perspective on bortezomib and second generation proteasome inhibitors versus future generation inhibitors of ubiquitin-proteasome system
Current Cancer Drug Targets 14:517–536.

https://doi.org/10.2174/1568009614666140804154511
- Google Scholar
(2006) Yeast UBL-UBA proteins have partially redundant functions in cell cycle control
Cell Division 1:28.

https://doi.org/10.1186/1747-1028-1-28
- Google Scholar
1. Enns GM
2. Shashi V
3. Bainbridge M
4. Gambello MJ
5. Zahir FR
6. Bast T
7. Crimian R
8. Schoch K
9. Platt J
10. Cox R
11. Bernstein JA
12. Scavina M
13. Walter RS
14. Bibb A
15. Jones M
16. Hegde M
17. Graham BH
18. Need AC
19. Oviedo A
20. Schaaf CP
21. Boyle S
22. Butte AJ
23. Chen R
24. Chen R
25. Clark MJ
26. Haraksingh R
27. Cowan TM
28. He P
29. Langlois S
30. Zoghbi HY
31. Snyder M
32. Gibbs RA
33. Freeze HH
34. Goldstein DB
35. FORGE Canada Consortium
(2014) Mutations in NGLY1 cause an inherited disorder of the endoplasmic reticulum-associated degradation pathway
Genetics in Medicine 16:751–758.

https://doi.org/10.1038/gim.2014.22
- Google Scholar
1. Fenteany G
2. Standaert RF
3. Lane WS
4. Choi S
5. Corey EJ
6. Schreiber SL
(1995) Inhibition of proteasome activities and subunit-specific amino-terminal threonine modification by lactacystin
Science 268:726–731.

https://doi.org/10.1126/science.7732382
- Google Scholar
(2002) Complementary whole-genome technologies reveal the cellular response to proteasome inhibition by PS-341
PNAS 99:1461–1466.

https://doi.org/10.1073/pnas.032516399
- Google Scholar
(2014) Random and targeted transgene insertion in Caenorhabditis elegans using a modified Mos1 transposon
Nature Methods 11:529–534.

https://doi.org/10.1038/nmeth.2889
- Google Scholar
1. Gauss R
2. Jarosch E
3. Sommer T
4. Hirsch C
(2006) A complex of Yos9p and the HRD ligase integrates endoplasmic reticulum quality control into the degradation machinery
Nature Cell Biology 8:849–854.

https://doi.org/10.1038/ncb1445
- Google Scholar
1. Gibson DG
2. Young L
3. Chuang RY
4. Venter JC
5. Hutchison CA
6. Smith HO
(2009) Enzymatic assembly of DNA molecules up to several hundred kilobases
Nature Methods 6:343–345.

https://doi.org/10.1038/nmeth.1318
- Google Scholar
(2013) Integration of the unfolded protein and oxidative stress responses through SKN-1/Nrf
PLoS Genetics 9:e1003701.

https://doi.org/10.1371/journal.pgen.1003701
- Google Scholar
1. Gomez TA
2. Kolawa N
3. Gee M
4. Sweredoski MJ
5. Deshaies RJ
(2011) Identification of a functional docking site in the Rpn1 LRR domain for the UBA-UBL domain protein Ddi1
BMC Biology 9:33.

https://doi.org/10.1186/1741-7007-9-33
- Google Scholar
1. Groll M
2. Schellenberg B
3. Bachmann AS
4. Archer CR
5. Huber R
6. Powell TK
7. Lindow S
8. Kaiser M
9. Dudler R
(2008) A plant pathogen virulence factor inhibits the eukaryotic proteasome by a novel mechanism
Nature 452:755–758.

https://doi.org/10.1038/nature06782
- Google Scholar
1. Hunt-Newbury R
2. Viveiros R
3. Johnsen R
4. Mah A
5. Anastas D
6. Fang L
7. Halfnight E
8. Lee D
9. Lin J
10. Lorch A
11. McKay S
12. Okada HM
13. Pan J
14. Schulz AK
15. Tu D
16. Wong K
17. Zhao Z
18. Alexeyenko A
19. Burglin T
20. Sonnhammer E
21. Schnabel R
22. Jones SJ
23. Marra MA
24. Baillie DL
25. Moerman DG
(2007) High-throughput in vivo analysis of gene expression in Caenorhabditis elegans
PLoS Biology 5:e237.

https://doi.org/10.1371/journal.pbio.0050237
- Google Scholar
1. Kamath RS
2. Ahringer J
(2003) Genome-wide RNAi screening in Caenorhabditis elegans
Methods 30:313–321.

https://doi.org/10.1016/S1046-2023(03)00050-1
- Google Scholar
1. Kaplun L
2. Tzirkin R
3. Bakhrat A
4. Shabek N
5. Ivantsiv Y
6. Raveh D
(2005) The DNA damage-inducible UbL-UbA protein Ddi1 participates in Mec1-mediated degradation of Ho endonuclease
Molecular and Cellular Biology 25:5355–5362.

https://doi.org/10.1128/MCB.25.13.5355-5362.2005
- Google Scholar
1. Keith SA
2. Maddux SK
3. Zhong Y
4. Chinchankar MN
5. Ferguson AA
6. Ghazi A
7. Fisher AL
(2016) Graded Proteasome Dysfunction in Caenorhabditis elegans Activates an Adaptive Response Involving the Conserved SKN-1 and ELT-2 Transcription Factors and the Autophagy-Lysosome Pathway
PLoS Genetics 12:e1005823.

https://doi.org/10.1371/journal.pgen.1005823
- Google Scholar
1. Kim I
2. Ahn J
3. Liu C
4. Tanabe K
5. Apodaca J
6. Suzuki T
7. Rao H
(2006) The Png1-Rad23 complex regulates glycoprotein turnover
Journal of Cell Biology 172:211–219.

https://doi.org/10.1083/jcb.200507149
- Google Scholar
(2007) Disease-associated prion protein oligomers inhibit the 26S proteasome
Molecular Cell 26:175–188.

https://doi.org/10.1016/j.molcel.2007.04.001
- Google Scholar
(2011) Specific SKN-1/Nrf stress responses to perturbations in translation elongation and proteasome activity
PLoS Genetics 7:e1002119.

https://doi.org/10.1371/journal.pgen.1002119
- Google Scholar
1. Meiners S
2. Heyken D
3. Weller A
4. Ludwig A
5. Stangl K
6. Kloetzel PM
7. Krüger E
(2003) Inhibition of proteasome activity induces concerted expression of proteasome genes and de novo formation of Mammalian proteasomes
Journal of Biological Chemistry 278:21517–21525.

https://doi.org/10.1074/jbc.M301032200
- Google Scholar
1. Melo JA
2. Ruvkun G
(2012) Inactivation of conserved C. elegans genes engages pathogen- and xenobiotic-associated defenses
Cell 149:452–466.

https://doi.org/10.1016/j.cell.2012.02.050
- Google Scholar
1. Meng L
2. Mohan R
3. Kwok BHB
4. Elofsson M
5. Sin N
6. Crews CM
(1999) Epoxomicin, a potent and selective proteasome inhibitor, exhibits in vivo antiinflammatory activity
PNAS 96:10403–10408.

https://doi.org/10.1073/pnas.96.18.10403
- Google Scholar
(2012) CloudMap: a cloud-based pipeline for analysis of mutant genome sequences
Genetics 192:1249–1269.

https://doi.org/10.1534/genetics.112.144204
- Google Scholar
1. Niu W
2. Lu ZJ
3. Zhong M
4. Sarov M
5. Murray JI
6. Brdlik CM
7. Janette J
8. Chen C
9. Alves P
10. Preston E
11. Slightham C
12. Jiang L
13. Hyman AA
14. Kim SK
15. Waterston RH
16. Gerstein M
17. Snyder M
18. Reinke V
(2011) Diverse transcription factor binding features revealed by genome-wide ChIP-seq in C. elegans
Genome Research 21:245–254.

https://doi.org/10.1101/gr.114587.110
- Google Scholar
1. Nowicka U
2. Zhang D
3. Walker O
4. Krutauz D
5. Castañeda CA
6. Chaturvedi A
7. Chen TY
8. Reis N
9. Glickman MH
10. Fushman D
(2015) DNA-damage-inducible 1 protein (Ddi1) contains an uncharacteristic ubiquitin-like domain that binds ubiquitin
Structure 23:542–557.

https://doi.org/10.1016/j.str.2015.01.010
- Google Scholar
1. Oliveira RP
2. Porter Abate J
3. Dilks K
4. Landis J
5. Ashraf J
6. Murphy CT
7. Blackwell TK
8. Abate JP
(2009) Condition-adapted stress and longevity gene regulation by Caenorhabditis elegans SKN-1/Nrf
Aging Cell 8:524–541.

https://doi.org/10.1111/j.1474-9726.2009.00501.x
- Google Scholar
1. Paek J
2. Lo JY
3. Narasimhan SD
4. Nguyen TN
5. Glover-Cutter K
6. Robida-Stubbs S
7. Suzuki T
8. Yamamoto M
9. Blackwell TK
10. Curran SP
(2012) Mitochondrial SKN-1/Nrf mediates a conserved starvation response
Cell Metabolism 16:526–537.

https://doi.org/10.1016/j.cmet.2012.09.007
- Google Scholar
1. Paix A
2. Wang Y
3. Smith HE
4. Lee CY
5. Calidas D
6. Lu T
7. Smith J
8. Schmidt H
9. Krause MW
10. Seydoux G
11. Lee C-YS
(2014) Scalable and versatile genome editing using linear DNAs with microhomology to Cas9 Sites in Caenorhabditis elegans
Genetics 198:1347–1356.

https://doi.org/10.1534/genetics.114.170423
- Google Scholar
1. Paul S
(2008) Dysfunction of the ubiquitin-proteasome system in multiple disease conditions: therapeutic approaches
BioEssays 30:1172–1184.

https://doi.org/10.1002/bies.20852
- Google Scholar
(2014) p97-dependent retrotranslocation and proteolytic processing govern formation of active Nrf1 upon proteasome inhibition
eLife 3:e01856.

https://doi.org/10.7554/eLife.01856
- Google Scholar
1. Radhakrishnan SK
2. Lee CS
3. Young P
4. Beskow A
5. Chan JY
6. Deshaies RJ
(2010) Transcription factor Nrf1 mediates the proteasome recovery pathway after proteasome inhibition in mammalian cells
Molecular Cell 38:17–28.

https://doi.org/10.1016/j.molcel.2010.02.029
- Google Scholar
1. Saez I
2. Vilchez D
(2014) The Mechanistic Links Between Proteasome Activity, Aging and Age-related Diseases
Current Genomics 15:38–51.

https://doi.org/10.2174/138920291501140306113344
- Google Scholar
1. Sha Z
2. Goldberg AL
(2014) Proteasome-mediated processing of Nrf1 is essential for coordinate induction of all proteasome subunits and p97
Current Biology 24:1573–1583.

https://doi.org/10.1016/j.cub.2014.06.004
- Google Scholar
1. Sirkis R
2. Gerst JE
3. Fass D
(2006) Ddi1, a eukaryotic protein with the retroviral protease fold
Journal of Molecular Biology 364:376–387.

https://doi.org/10.1016/j.jmb.2006.08.086
- Google Scholar
(2011) Road to ruin: targeting proteins for degradation in the endoplasmic reticulum
Science 334:1086–1090.

https://doi.org/10.1126/science.1209235
- Google Scholar
1. Snyder H
2. Mensah K
3. Theisler C
4. Lee J
5. Matouschek A
6. Wolozin B
(2003) Aggregated and monomeric alpha-synuclein bind to the S6' proteasomal protein and inhibit proteasomal function
Journal of Biological Chemistry 278:11753–11759.

https://doi.org/10.1074/jbc.M208641200
- Google Scholar
(2014) Regulation of synaptic nlg-1/neuroligin abundance by the skn-1/Nrf stress response pathway protects against oxidative stress
PLoS Genetics 10:e1004100.

https://doi.org/10.1371/journal.pgen.1004100
- Google Scholar
1. Steffen J
2. Seeger M
3. Koch A
4. Krüger E
(2010) Proteasomal degradation is transcriptionally controlled by TCF11 via an ERAD-dependent feedback loop
Molecular Cell 40:147–158.

https://doi.org/10.1016/j.molcel.2010.09.012
- Google Scholar
(2015) Lipid-mediated regulation of SKN-1/Nrf in response to germ cell absence
eLife 4:e07836.

https://doi.org/10.7554/eLife.07836
- Google Scholar
(2016) The cytoplasmic peptide:N-glycanase (NGLY1) - Structure, expression and cellular functions
Gene 577:1–7.

https://doi.org/10.1016/j.gene.2015.11.021
- Google Scholar
1. Taylor RC
2. Dillin A
(2011) Aging as an event of proteostasis collapse
Cold Spring Harbor Perspectives in Biology 3:a004440.

https://doi.org/10.1101/cshperspect.a004440
- Google Scholar
1. Tomko RJ
2. Hochstrasser M
(2013) Molecular architecture and assembly of the eukaryotic proteasome
Annual Review of Biochemistry 82:415–445.

https://doi.org/10.1146/annurev-biochem-060410-150257
- Google Scholar
1. Wang W
2. Chan JY
(2006) Nrf1 is targeted to the endoplasmic reticulum membrane by an N-terminal transmembrane domain. Inhibition of nuclear translocation and transacting function
Journal of Biological Chemistry 281:19676–19687.

https://doi.org/10.1074/jbc.M602802200
- Google Scholar
1. Ward JD
(2015) Rapid and precise engineering of the Caenorhabditis elegans genome with lethal mutation co-conversion and inactivation of NHEJ repair
Genetics 199:363–377.

https://doi.org/10.1534/genetics.114.172361
- Google Scholar
1. Wen C
2. Greenwald I
(1999) p24 proteins and quality control of LIN-12 and GLP-1 trafficking in Caenorhabditis elegans
Journal of Cell Biology 145:1165–1175.

https://doi.org/10.1083/jcb.145.6.1165
- Google Scholar
1. White RE
2. Dickinson JR
3. Semple CA
4. Powell DJ
5. Berry C
(2011) The retroviral proteinase active site and the N-terminus of Ddi1 are required for repression of protein secretion
FEBS Letters 585:139–142.

https://doi.org/10.1016/j.febslet.2010.11.026
- Google Scholar
1. Wójcik C
2. DeMartino GN
(2002) Analysis of Drosophila 26 S proteasome using RNA interference
Journal of Biological Chemistry 277:6188–6197.

https://doi.org/10.1074/jbc.M109996200
- Google Scholar
1. Zhang Y
2. Hayes JD
(2013) The membrane-topogenic vectorial behaviour of Nrf1 controls its post-translational modification and transactivation activity
Scientific Reports 3:2006.

https://doi.org/10.1038/srep02006
- Google Scholar
1. Zhang Y
2. Li S
3. Xiang Y
4. Qiu L
5. Zhao H
6. Hayes JD
(2015) The selective post-translational processing of transcription factor Nrf1 yields distinct isoforms that dictate its ability to differentially regulate gene expression
Scientific Reports 5:12983.

https://doi.org/10.1038/srep12983
- Google Scholar
1. Zhang Y
2. Lucocq JM
3. Yamamoto M
4. Hayes JD
(2007) The NHB1 (N-terminal homology box 1) sequence in transcription factor Nrf1 is required to anchor it to the endoplasmic reticulum and also to enable its asparagine-glycosylation
Biochemical Journal 408:161–172.

https://doi.org/10.1042/BJ20070761
- Google Scholar
1. Zhang Y
2. Ren Y
3. Li S
4. Hayes JD
(2014) Transcription factor Nrf1 is topologically repartitioned across membranes to enable target gene transactivation through its acidic glucose-responsive domains
PLoS ONE 9:e93458.

https://doi.org/10.1371/journal.pone.0093458
- Google Scholar

Article and author information

Author details

Nicolas J Lehrbach
1. Department of Molecular Biology, Massachusetts General Hospital, Boston, United States
2. Department of Genetics, Harvard Medical School, Boston, United States
Contribution
NJL, Conception and design, Acquisition of data, Analysis and interpretation of data, Drafting or revising the article

Competing interests
The authors declare that no competing interests exist.
Gary Ruvkun
1. Department of Molecular Biology, Massachusetts General Hospital, Boston, United States
2. Department of Genetics, Harvard Medical School, Boston, United States
Contribution
GR, Conception and design, Analysis and interpretation of data, Drafting or revising the article

For correspondence
ruvkun@molbio.mgh.harvard.edu

Competing interests
The authors declare that no competing interests exist.

"This ORCID iD identifies the author of this article:" 0000-0002-7473-8484

Funding

National Institutes of Health (R01 AG016636)

Gary Ruvkun

Human Frontier Science Program (Long Term Fellowship)

Nicolas J Lehrbach

The funders had no role in study design, data collection and interpretation, or the decision to submit the work for publication.

Acknowledgements

We would like to thank Ulandt Kim for high-throughput sequencing. We thank Fei Ji and Ruslan Sadreyev for bioinformatics advice. We thank Fred Goldberg and Keith Blackwell and members of their labs for useful discussions. This work was supported by an NIH R01 grant to GR [R01 AG016636]. NL was supported by a Human Frontier Science Program Postdoctoral Fellowship.

Copyright

This article is distributed under the terms of the Creative Commons Attribution License, which permits unrestricted use and redistribution provided that the original author and source are credited.